Puzzle design case study: LSL1 (SPOILERS)

[adeyke]

This recent thread reminded me of this less recent thread, where I complained about LSL1's puzzles. So I thought I'd do a more in-depth analysis of those.

One thing that really influenced my view of puzzle design is Ron Gilbert's Why Adventure Games Suck, especially the section about backwards puzzles. I'm going to use the terminology "problem-first puzzle" and "solution-first puzzle". The problem with solution-first puzzles is that it leads to the behavior of first just picking up everything everywhere, with the logic that if the game allows you to get something, you'll probably need it. The puzzle itself is then reduced to just looking through your keyring for the right key.

Problem-first puzzles have much more potential for that satisfying "a-ha" moment. You'd first encounter a problem, then try to work your way past it. There are then two possibilities. You could passively encounter the solution, at which point the puzzle is to make the mental connection between the two (Gilbert gives the example of finding a rope and then concluding that you could use it to climb down a crevice). Or you could hypothesize a potential solution and actively work towards solving it (e.g. breaking into a toolshed in order to look for a rope). I think active problem-first puzzles are the most satisfying. For both solution-first puzzles and passive problem-first puzzles, it all depends on how much thought the connection requires. Figuring out that a disco passcard lets you into the disco doesn't make the player feel very clever.

I'm going to base this analysis on the AGI version of LSL1. And I'm just going to discuss what I see as puzzles, so simple things like "open the door in order to go through it" are omitted, as are any entirely optional actions (even if they give points). Also, while I noted it in the post title, I'll warn again that this has all the spoilers.

End objective
As the manual puts it, "to find and seduce the girl of your dreams".

This is already a problem. You meet several women in the game, but the actual endgame comes from going up to the penthouse. You're never told who is up there or given an indication that she'd appreciate you intruding like that. You're apparently only going up there because you can. Finding Eve is accidental.

Puzzles

Problem: You need a password for the door in Lefty's.
Solution: Read the bathroom graffiti.

This one could go either way. It's very possible that you'll find the password before learning you need it (solution-first). Or you'll figure out that you need the password and then stumble upon it (passive problem-first). It's also possible find out you need a password and then specifically look for it in the graffiti (active problem-first), since that's not an absurd place for it.

Problem: You need to distract the pimp.
Solution: Give the drunk in the hallway a drink to get a remote control.

This is likely solution-first (you'll pass the drunk on the way to get the password), and it's entirely passive. If you were just trying to find a remote control, there's no logical reason to expect it from that drunk. If, on the other hand, you already have the remote control, the puzzle solves itself.

Problem: You can't get into the disco.
Solution: Get the passcard from the ashtray in the casino.

This is either solution-first or passive problem-first. You wouldn't specifically look for the passcard there, but once you have it, its use is obvious.

Problem: Fawn wants things.
Solution: Give her the rose from Lefty's hallway, the ring from the sink, and the candy from upstairs.

This is also solution-first. She doesn't ever say that she wants these items, so you never really have that as an objective. You're just expected to already have them (which is likely, since you start the game at Lefty's) and to try giving them to her. Even if you do speak with her first and decide to try to get gifts for her, it's still passive; it doesn't make sense to specifically search a bar looking for gifts.

Problem: Fawn wants wine.
Solution: Use the radio, wait for a commercial, then go to the payphone and call that number.

This is problem-first, and it's probably also active. Having something relevant show up immediately on turning on a radio or tv doesn't really work in real life, but it is a common trope in fiction. This one gets a bonus point for having another source of wine that doesn't work but it then loses that point because you only find out that other wine doesn't work from unexpectedly dying.

Problem: Fawn ties you up.
Solution: Give the wandering drunk wine in order to get a knife.

Here, you either already have the solution, or you're stuck. So it's either solution-first or an unexpected game over.

Problem: Faith wants Spanish fly.
Solution: Get it from the window behind Lefty's.

This is, since the narrator tells you to get some medical stimulants. And it's active, in that you see them in the window and try to work towards getting them.

Problem: You need a rope to tie to the fire escape.
Solution: Use the one Fawn tied you to the bed with.

This is probably solution-first, and a dead end if you failed to pick it up. If, on the other hand, you start out trying to get the pills, you'd have to do the whole Fawn subplot first in order to get that rope, making for an absurd passive problem-first puzzle. I do have to give props to the game for hinting at the need for rope via the magazine.

Problem: You need a hammer to smash the window.
Solution: Search the dumpster.

This could be solution-first or problem-first. However, it's very likely passive; you might search the dumpster because you're expecting to find something useful there, but you wouldn't go there for a hammer specifically.

Problem: Eve wants an apple.
Solution: Buy one from the wandering barrel man.

This is almost certainly solution-first. It would be really weird to get to the penthouse and then go back outside to look for an apple, especially since you're never given that as objective (you're just supposed to guess it from her name).

Conclusion

When I said LSL1 didn't have good puzzles, this is what I meant. There aren't a lot of puzzles in the first place. In most cases, the puzzle is solved through adventure game kleptomania rather than through specifically working out a solution. This means you spend your time either at the point where the solution is obvious (you already have the item needed) or where you're completely in the dark about it, without the middle ground of having some inkling and working towards solving it. And it means that you spend a lot of time looking at things at the meta level (things like "what items look like they can be picked up?" or "if it's possible to do this, there must be a reason to do it") rather than the in-universe level ("if I were in this situation, what would I do?").

I'm not sure if this analysis is helpful to anyone else, but I found it useful to crystalize my thoughts (I hadn't really thought about the passive/active distinction before). If people are interested, I could try to do this sort of analysis on other games in the future. If not, that's also okay.

[BBP]

First off, my apologies for not reading the text you link to - I have an aversion to people using the word "suck" when something is not related to suction, particularly in a title. So I'm going with reading your post. I'm interested in where your aversion for these puzzles come from, especially in such a trailblazing game as LSL1. It's the core of Sierra you're up against.

I wouldn't necessarily make the partition between active and passive. Adventure gaming simulates a warped version of real life, you could argue, and a lot of the objects you mention, like LSL1's rope, the hammer in the garbage, the disco pass, the rose and the candy - I know it's a bit of a scrounging mind but one could argue that these things are convenient to have just in case. Especially since you're dumped somewhere with a vague objective and too little money to go through the night. You refer to this as a "problem" but I don't see how it is a problem. A piece of string or a screwdriver (adventure game stable item no 1,especially in Phantas 2) is just the thing a self-respecting adventurer would want, and nobody would leave a diamond ring by the sink or a pass in the ashtray either.

So here's the core of what I don't get about your analysis:

Why do you think it's a problem to have an item that forms the solution to the puzzle before you reach the puzzle?


And what's more, what's the alternative? A game like LSL3? Where you spend aeons on the weight machines and can't visit the show until you've picked up the towel?

I'll agree that the disco pass is not too difficult a puzzle, but one could argue that the disco is an essential part of the landscape and making access difficult could discourage players. Although in a respiffed update I'd probaby add a puzzle similar to the Photo ID-one in LSL7. Maybe with less p0rn.

I forgot how Softporn handled the disco pass...

[adeyke]

To illustrate the difference, consider these scenarios:
1. On one screen, you see a stick lying around. A few screens later, you see a dog you need to distract.
2. On one screen, you see a dog you need to distract. A few screens later, you see a stick lying around.
3. On one screen, you see a dog you need to distract. A few screens later, you see a dead tree from which you can snap off a branch.

These all end up with you throwing the stick for the dog, but the thoughts and feelings involved differ a lot.

In the first scenario, you'd probably pick up the stick just because it's there. Maybe you can tell it's a sprite because it doesn't blend into the background, or maybe the screen description points it out. Either way, you're just picking it up because that's how adventure games work, not because you actually have any plans for it or because you think your character would be inclined to pick up random sticks. Then, when you encounter the dog, you already have the stick, so all you had to do to solve the puzzle is look in your inventory.

In the second scenario, you first encounter the dog. You now have a goal of trying to see what's past the dog, and something you're specifically looking for in order reach that goal. You keep this in your mind as you explore further. Later, when you find the stick, you go "A-ha! This is just what I need." The character now has a real motivation for taking the stick, beyond being controlled by an adventure game player.

In the third scenario, this is much the same, but there's an additional step. You go from thinking "I need something to distract the dog" to thinking "I could throw a stick for the dog" to thinking "I could a stick from that tree". Instead of just keeping an eye out for the solution in case it falls into your lap, you're now actively doing something in order to reach that solution.

It all comes down to the purpose of the puzzles. They're there to challenge and engage the player, to make them think, and to make them feel clever for solving them. If, on first reaching the pimp, the player already has the remote control, there's no puzzle left anymore. They don't feel clever for knowing that a remote control can be used on a tv. On the other hand, if the player first encounters the pimp and sees the tv, they might have the insight that they need to find a remote control, but they don't have any means of acting on that insight. There's never a point where they could say "I'm looking for a remote control and I'm doing this in order to get one". When they do get the remote from the drunk, it'll be a surprise to them. They never get the chance to feel clever about finding it.

I don't have any suggestions for fixing the puzzles now. Things like this need to be considered early on in the design; trying to retrofit it into a finished game doesn't work well. When planning the puzzles of an adventure game, I think effort should be made to present the problems prior to their solutions and to place the solutions somewhere the player would expect them. I know puzzle design can be tricky; it has to avoid both the extremes of being too obvious and of being too obscure. I just think that, for much of LSL1, the puzzles didn't fall within that sweet spot: you either solve them without needing to think, or you can't solve them no matter how much you think.

And I do acknowledge that LSL1 was hamstrung by needing to adhere to Softporn's design. They fortunately changed some of the puzzles, but most of them are still the same. For reference, the disco passcard is also found in an ashtray there. So yes, I can blame Softporn for that puzzle, but that's just a reason why it's bad; it doesn't make it good.

[Collector]

I'm not sure if this analysis is helpful to anyone else, but I found it useful to crystalize my thoughts (I hadn't really thought about the passive/active distinction before). If people are interested, I could try to do this sort of analysis on other games in the future. If not, that's also okay.

Why not. At least it stimulates discussion.

In your last post you give three examples that while the player's perception in each one may be different to control the sequence of events would probably make the game too linear. I suppose that you could do something like program it to not show the stick until a "dog seen" flag had been set, but for other scenarios it may be harder to make a convincing method to control the sequence of events or might not be possible. However, when you can it could make a game feel more organic.

Some of this problem with inventory is from the early established adventure game standard of pickup everything not nailed down. I guess you could make games that ignore this by having many more inventory items that can be acquired than what is needed by the game. Add in multiple solutions for inventory puzzles of varying success. Maybe impose a limit to the number that you can carry at any one point in time. In other words add some cost to adding an inventory item so the player will only take what he thinks will be useful at the time.

It is a nice thought exercise.

[BBP]

It would make me very angry if I had to see the dog before the stick would appear. I begin my adventures with exploring the territory and even if there is a reason for the part to be late, like in KQ3 with the eagle feather (hears some cursing from players who never got to see the eagle) - or with the infernal towel from LSL3 - this is just fake difficulty.

The puzzle you describe is in KQ5 - twice you need to distract an animal and there are two objects you can use that are interchangeable - there's a chance you'll encounter one of these before picking up the item to cast, but the other will pop up when you enter that screen with the appropriate flag down.
However, the second one will only happen once, and you'll have to be quick about it.
The puzzle combination has some fairness about it but with the one-time-event (game can't be finished if you miss it and you won't know) in place the net result is it's too frustrating.

Part of the puzzle, the breaking off of the twig, is in LSL3 - which I think is a dreadful game because it's so manual-based. You won't find out you'll need the twig until you've read the manual.

I don't know who "the player" is you refer to, but it is not about me. I want fair play, not this backhanded fake-out "Oh look, a stick has magically appeared now that you need it!"

What are some successful puzzles, and how do they rank up against this article?
The best puzzle winner of the poll here is Le Serpent Rouge - which cannot fit in this simple division suggested.

[adeyke]

I'm not sure if this analysis is helpful to anyone else, but I found it useful to crystalize my thoughts (I hadn't really thought about the passive/active distinction before). If people are interested, I could try to do this sort of analysis on other games in the future. If not, that's also okay.

Why not. At least it stimulates discussion.

In your last post you give three examples that while the player's perception in each one may be different to control the sequence of events would probably make the game too linear. I suppose that you could do something like program it to not show the stick until a "dog seen" flag had been set, but for other scenarios it may be harder to make a convincing method to control the sequence of events or might not be possible. However, when you can it could make a game feel more organic.

The issue of linearity is a valid one. If you allow the player a lot of freedom, you no longer necessarily control the order in which they encounter things. The best you can do then is to try to make it more likely that they'll encounter the problem first. For example, if a player can either go left to the stick or right to the dog, but they also have an objective pointing them to the right, they'll probably go right first.

Of course, there are plenty of cases where (sections of) games are completely linear and you always encounter the solution before the problem. Those don't have the excuse of nonlinearity. So the most obvious solution there is to just to adjust the geography: place the screens in such a way that you encounter the problem first.

Another possibility is to make thought required in getting the solution. That would go to something like the dead tree. I didn't list it as one of the scenarios, but the player could also encounter the dead tree first and then the dog. In that case, they might first disregard the tree as just part of the scenery and then only go back to interact with it later. Or applying the solution could require thought. If the player already has it in their inventory but its use isn't obvious, there's still room for player cleverness.

I definitely wouldn't approve of the stick just appearing after you've encountered the dog. That violates another of Gilbert's Rules of Thumb: unconnected events. If the player already searched a screen and saw no stick, and if there's no logical causal connection between seeing the dog and a stick appearing, they have no reason to go back to that screen and expect anything to have changed.

Another possibility would be to have the stick there and use the knowledge flag to determine if the player can pick it up or not. But I have very mixed feelings about such knowledge flags. They work well when neither the character nor the player has any real motivation for an action (i.e. the player is just clicking random things). However, they can be very frustrating if either the player knows they need to do something (e.g. because they've played the game before) or if it's something the character actually has some reason to do (some items do just look like they'll be useful to carry around, even without a specific use in mind). It then becomes a case of "how can I convince my character that they actually need to do this?".

Some of this problem with inventory is from the early established adventure game standard of pickup everything not nailed down. I guess you could make games that ignore this by having many more inventory items that can be acquired than what is needed by the game. Add in multiple solutions for inventory puzzles of varying success. Maybe impose a limit to the number that you can carry at any one point in time. In other words add some cost to adding an inventory item so the player will only take what he thinks will be useful at the time.

It is a nice thought exercise.

It's an interesting idea but also a dangerous one.

If there are just a few no use, it wouldn't really change player behavior. They'd still pick everything up; they'd just leave the game wondering what those items were for. In order to really discourage the player from thinking that being able to pick something up is automatically progress, you'd have to include a lot of extra items. And this brings its own set of problems.

The big one is the combinatorial explosion. The more items there are, the more possible interactions there are. Some of the "wrong" solutions might also be sensible. In that case, the game would either need to also allow that (which requires the game designer to think through all of them and make sure that it doesn't allow any sequence breaking) or give a meaningful explanation for why that particular solution doesn't work. For example, if the player needs to carry some water in a bottle, they might instead try to use the red herring mug they're carrying around. If the game just gives a generic "you can't do that" message, they might conclude that it's not possible to get the water. So the game would either need to allow the mug as alternative or give an explanation for why the mug doesn't work for that (e.g. it's cracked) but something else would.

As for finite inventory, I'm really not a fan. That has the potential of just busy work: instead of picking up item X and going to the location where it's needed, they might need to drop item Y, pick up item X, go to the location where it's needed, then go back to pick up item Y. And if any screens ever become inaccessible over the course of the game, there's a big risk of a dead end if the player left an item there. And if the player can drop items anywhere, they'll probably end up somewhere illogical; the player might need to remember to go back to the kitchen to find the branch and into the forest to find the bowl, for example.

I should stress that the list is "Gilbert's Rules of Thumb", not "Gilbert's list of things that make games irredeemable trash if they're violated". The idea is that following the rules will make the game more satisfying to play and will facilitate the suspension of disbelief, but it's very easy to violate them accidentally (good adventure game design is hard).

I think when it comes to LSL1 in particular, the active vs. passive distinction is a lot more significant than the problem-first vs. solution-first distinction. Seeing the problem first can give the player and objective and something to work towards, but that doesn't help if there isn't actually anything they can to help them reach that objective. Too often, the motivation for the next action is just "because it's there" rather than because the player thinks it will help them get closer to their goals, and too often, the player will achieve those goals just by accident. That brings me to an idea to help improve a design: write a walkthrough of the game, annotating each action with "because of [something]" or "in order to [do something]". If there is no sensible reason for a particular action, or if there aren't any actions actually motivated by one of the game's sub-goals, that may be something to address.

[Tawmis]

End objective
As the manual puts it, "to find and seduce the girl of your dreams".
This is already a problem. You meet several women in the game, but the actual endgame comes from going up to the penthouse. You're never told who is up there or given an indication that she'd appreciate you intruding like that. You're apparently only going up there because you can. Finding Eve is accidental.

Feeling a strong connection to Mr. Laffer, as I do, I think you may be reading way too much into this.

My impression is that, pretty much everything Larry does (good or bad) is accidental. (Sort of like Roger Wilco). Larry is an older guy, who hasn't not had the pleasure of having sex with a woman. So, you'd think sex with just about any woman is going to be the "girl of his dreams." And we get the shot at the hooker, as Larry, but that immediately tells you that it wasn't satisfying.

But just like any of the adventure games, you're going into areas "just because you can."

It's not an open world, so you have to open "doors" to get to the next area, more often than not. So in order for Larry to accidentally get to Eve, he has to open "other doors" to get there.

When you met your boyfriend/girlfriend, I doubt that you went there knowing you would. It just happened. Whether you were at a friend's party, at a bar, met them online - you pretty much accidentally stumble into them and find that they're the one for you. (Well, at least for the first game - and like life, you may "find your one true love" who turns out - a few years later, after you've gotten to know them - that they're not all they're cracked up to be).

[adeyke]

I get what you're saying, but I think there's a huge qualitative difference between meeting someone in a party, a bar, or online vs. meeting someone because you distracted their guard, entered their hotel suite without permission, and then saw them bathing. That ending would have worked a lot better if Eve had met Larry earlier and invited him in, but then Faith just didn't get the memo or wouldn't cooperate for some reason.

The bit about objectives is really about player agency. There's a bigger sense of agency if you have some of sort of goal that you feel your actions are working towards. Quest for Glory 1 is an example of a Sierra game that did that well. You start out with the quest to deal with the brigand problem, and soon also get the quests to return the Baron's children and deal with the brigand warlock, as well some smaller sidequests. Then you spend a lot of time doing those sidequests, talking to people to learn more about your quest, finding out about the curse, and gathering the ingredients for the dispel potion. And you spend time earning gold for spells/equipment/potions and improving your stats in order to be ready for that final confrontation. Almost everything you do follows from you trying to do your quest or from you trying to become a hero. When you finally succeed, you can really feel you've earned it.

Without clear objectives, there's the risk of players just getting tunnel vision and thinking only in terms of "how to I leave this screen" and "how do I solve this puzzle", without realizing what those actually mean in the big picture.

I do get your point about Larry's adventures just being accidental. LSL2 had a big joke about that with the onklunk. While I think that resulted in really bad gameplay (there were lots and lots of unexpected deaths), I do admit the joke itself was funny. However, even if the poor narrative cohesion and puzzle design were a deliberate decision in order to say something about the protagonist, you're still left with poor narrative cohesion and puzzle design.

[Tawmis]

Without clear objectives, there's the risk of players just getting tunnel vision and thinking only in terms of "how to I leave this screen" and "how do I solve this puzzle", without realizing what those actually mean in the big picture.

I definitely see where you're coming from. For me, I think Larry has always had the "simplest" goal/end game.

Games like Quest for Glory, with multiple ways to beat it (depending on your "class"), or the more, story driven, heavily detailed Gabriel Knight, or even the "detail oriented" Police Quest; make those games more difficult. Leisure Suit Larry, is at least, pretty simple (with both puzzles and story), so it's the easiest to keep moving forward. Police Quest can take a turn because you didn't read someone their rights, or search their body; Gabriel Knight, you have to really pay attention to all of the details. Quest for Glory, steps outside of the "realistic" world and makes it more complicated. For me, I struggle more with other Sierra games, then I do with Larry games. So, for myself, I don't get stuck on the "how do I leave this screen" for Larry; where as GK and PQ have more often than not stumped me; and this is with going back and doing replays of the game. Even King's Quest to some degree; because the earlier games pretty much required knowing a lot of lore, fairy tails, and kid's stories, to understand what to do with some of the characters you interact with.

But like I said, I see where you're coming from. And definitely, in hindsight, virtually any game can be made better. Even if it was Eve's limo driving by Larry at the intro while he's outside Lefty's and sees her for the briefest of moments.